This study aims to: (1) Describe the type and extent of changes in nutritional status that develop across the spectrum of HIV infection during its treatment, including etiology of nutritional changes and adequacy of nutritional intake; changes in body composition by weight and other anthropometric measures and bioelectrical impedance analysis (BIA); biochemical parameters associated with nutritional status or suggested to be associated with HIV infection (cholesterol, triglycerides, vitamin B12, and folate); and (2) Explore the extent to which nutritional status serves as a cofactor to impaired immune function during HIV infection and its treatment. The effect of timing and pattern of collecting dietary intake information on diversity of reported intake and dependability of reporting will also be determined. Initiated in July 1990, the study employs an initial cross-sectional design, comparing different HIV disease severity groups determined by CD4 count and HIV treatment protocols, combined with a longitudinal design, involving follow-up of subjects at monthly intervals for the first eight months and then at decreasing intervals. Recruitment was completed in May 1992, including 120 subjects of whom 13 were females. In response to early observations regarding the absence of anticipated measurable levels of Tumor Necrosis Factor (TNF), two substudies were conducted in collaboration with LIR/NIAID. In addition, two other substudies were conducted. The first evaluated the consequences of the observed inability of subjects to not eat or drink prior to BIA on the accuracy of body composition estimates by BIA. The second measures energy consumption by indirect calorimetry in order to validate estimates of energy requirements with these more precise indirect colorimetry measures.